Wireless Sensor Network (WSN), is a new information paradigm based on the collaboration of a large number of self-organized sensing nodes. With the increasing demand of cyber-physical interaction, wireless sensor networks have emerged as one of key technologies for many promising applications such as assisted living, military surveillance, infrastructure protection and scientific exploration. Sensor networks are acclaimed to be low-cost, low-profile, and easy to deploy. These attractive advantages, however, imply the resources available to individual nodes are severely limited. Although it is highly possible that the constraints on computation and storage disappear along with the fast development of fabrication techniques, the energy constraint is unlikely to fade away quickly. On the other hand, many sensor network based applications require a lifetime that ranges from several months to tens of years. In order to bridge the gap between limited energy supply and long-term operation requirement of applications, we then have to build extremely low duty-cycle sensor networks where during the operation of sensor applications, sensor nodes activate very briefly and stay in a dormant state for a very long period of time. In this dissertation research, we initiate the first systematic research for low-duty-cycle sensor networks, including a generic sensing architecture, a novel data forwarding scheme for intermittently connected networks and an energy synchronized communication middleware for energy-harvesting sensor networks. The goal of this dissertation research is to provide better understanding of how to build practical and efficient extremely low duty-cycle sensor networks and support those long-term applications such as structure monitoring, traffic control and so on. We hope, toward the very end, this dissertation research can assist the transition of sensor network technology from a research concept to a general-purpose technology available for use for a wide variety of research, government and industry purposes.